The Individualized Powerhouse: Mitofusin-2 Regulates Nucleus Accumbens Mitochondrial Influence on Individual Differences in Trait Anxiety

The vast range of human emotional experiences represents broad individuality and robust personality distinction. Experiencing a balanced emotional range supports general well-being, but some individuals are prone to higher levels of baseline anxiety and neuroticism, which can in turn become vulnerabilities to depression or other mood disorders. Understanding if there is an underlying molecular signature of high trait anxiety can lead to insight into how individual genetic, epigenetic, or developmental differences can lead to variance in mood disorder susceptibility or can inspire individualized treatments for those with mood disorders. In the current issue of Biological Psychiatry, Gebara et al. (1) explore the relationship between individual differences in anxiety-like behaviors and both the expression of mitochondrial genes and mitochondrial function in the nucleus accumbens (NAc).

Anxiety and depression are linked to functional changes in brain reward circuits, specifically in the NAc (2). NAc medium spiny neurons (MSNs) are the primary projection outputs from this region. Changes in their neuronal structure, excitability, and output are linked to changes in most reward-related behaviors, including those related to anxiety and depression in humans (2). NAc is a central node for individual differences in gene expression associated with susceptibility or resilience to chronic stress in mice (3). MSNs in the NAc are dynamically plastic in response to reward learning or to stress, increasing or decreasing dendritic complexity and dendritic spine density to support changes in neuronal excitability or output (2). The dependence of such dynamic remodeling on mitochondrial function, mitochondrial-dependent Ca²⁺ buffering, and available cellular energy is rapidly being explored in various neuronal subtypes, such as NAc MSNs. Gebara et al. (1) contribute to the small but growing literature examining mitochondrial function in the context of anxiety and mood disorders (4).

To study the molecular underpinnings of individual differences in baseline anxiety in laboratory rodents, the authors first defined their population of high-anxiety (HA) and low-anxiety (LA) male Wistar rats based on behavior in an elevated plus maze test. HA animals were defined as those with ≤5% open-arm duration (~8% of animals tested) and LA animals with ≥20% open-arm duration (~20% of animals tested), representing the naturally occurring extremes of observable anxiety-like behavior. Having defined HA and LA rats, Gebara et al. (1) observed significant differences between these groups in another measure of novelty-seeking and exploration, a novel object recognition task. Further, HA rats displayed motivational deficits associated with depression-like behavior. The elevated plus maze test is considered an anxiety test, but reduced open-arm exploration could also be indicative of broader reductions in reward-motivated exploration. As the behavioral profiles of HA rats are correlated across multiple motivation-related tasks, the defined “HA” phenotype may not specifically be related to anxiety and may represent an underlying reduced motivational state.

Functional and neuroanatomical variance in NAc MSNs is linked to anxiety- and depression-like behaviors in multiple behavioral models. Indeed, exposure to drugs of abuse leads to alterations in MSN dendritic spine formation and dendritic complexity (5). Conversely, various stressors and depression models induce dendritic atrophy and alter spine number in MSNs and other neurons in excitatory brain structures (2,6). Consistent with these findings, HA rats show reduced dendritic complexity, fewer dendritic spines, and fewer miniature excitatory postsynaptic potentials in NAc shell MSNs compared with LA rats. These changes were selective to the NAc, as dendritic arborization was comparable between HA and LA rats in the basal lateral amygdala, another region associated with disrupted dendritic complexity in the context of stress-related behaviors (6).

Gebara et al. (1) begin to parse the molecular underpinnings of this structure–function relationship by examining MSN mitochondria in the NAc. Mitochondrial function supports dendritic spine formation, maturation, and plasticity (7,8). In addition, mitochondrial morphology is disrupted in NAc D1-receptor–containing MSNs in behavioral motivation for a psychostimulant (9). Thus, it is not surprising that mitochondria from HA rats have reduced function, measured as lower complex I activity, and lower maximal respiration compared with LA rats. Anatomically, electron micrographs reveal equivalent mitochondrial numbers between HA and LA rats, but mitochondria from HA rats are larger and cover a greater cellular area, suggestive of swelling. Despite the larger size and increased mitochondria–mitochondria contact points between these mitochondria, contact points with the endoplasmic reticulum (ER) are reduced in HA rats.

These differences in mitochondrial morphology signal a potential variance in molecules that regulate mitochondrial fission, fusion, and overall morphology. Indeed, Gebara et al. (1) examined the expression levels of both messenger RNA and protein of fusion- and fission-related molecules. Of these molecules, HA rats had decreased expression uniquely of MFN2 messenger RNA and protein in the NAc. Decreased MFN2 expression is consistent with reduced ER contacts, as MFN2, more than its counterpart MFN1, specifically regulates ER–mitochondrial contacts, which are critical for efficient mitochondrial Ca²⁺ uptake (10). Interestingly, this decrease in MFN2 expression was observed in both D1- and D2-receptor–containing MSNs of the NAc shell, but not in D1/D2-negative cells in this region, or in cells of the cortex or the amygdala.

Gebara et al. (1) elegantly demonstrate the sufficiency of NAc MFN2 to regulate behavior by exogenously over-expressing MFN2 in the NAc in HA rats (MFN2-OE). This MFN2-OE significantly elevated NAc MFN2 in experimental rats, making MFN2 messenger RNA and protein levels in HA-MFN2-OE rats comparable to levels of sham-operated LA rats. Previously observed deficits in mitochondrial respirometry and morphology were increased to LA levels by MFN2-OE. At the cellular level, miniature excitatory postsynaptic potentials frequency also increased to LA levels, with reduced miniature excitatory postsynaptic potential amplitude, suggesting an increase in new dendritic spines. Further dendritic complexity increased to LA neuron complexity. Behaviorally, MFN2-OE was sufficient to increase elevated plus maze test open-arm time and saccharine preference, while decreasing forced swim immobility time, creating behavioral profiles that were indistinguishable from those of LA rats.

Taken together, these findings demonstrate natural variation in anxiety-like behavior, mitochondrial function, and MFN2 expression in the NAc. They establish that overexpression of MFN2 in HA rats can shift mitochondrial function, neuronal excitability, and anxiety-like behaviors to a LA phenotype. The variability of MFN2 expression and the resulting difference in mitochondrial morphology and function observed in HA versus LA rats presents a novel mechanism for naturally occurring individual differences in anxiety-like behavior. In addition, this work expands our understanding of the role of MFN2 in a new neuronal subpopulation, MSNs. Reduced or lowered MFN2 expression has variable outcomes depending on which cell population is involved. The larger, swollen mitochondria with fewer ER contact points observed here in NAc MSNs with lower MFN2 expression is comparable to phenotypes observed in pro-opiomelanocortin neurons of the arcuate nucleus of the hypothalamus when MFN2 is reduced or deleted (10). Interestingly, in this study, D1- and D2-MSNs show similar reductions in MFN2. How this comparable reduction influences known differences in baseline MSN subtype reactivity and event-related plasticity will inform our understanding of NAc MSNs.

Appropriate MFN2 function is critical for maintaining a robust fused mitochondrial network, in addition to regulating mitophagy and the normal cycle of degradation of damaged mitochondria. While MFN2 is a GTPase that mediates fusion events between mitochondrial outer membranes, mitochondrial–ER tethering points, mediated by MFN2, actually promote mitochondrial cinching and fission events. Fission can isolate damaged mitochondria and facilitate mitophagy, but in neurons, fission is also critical for trafficking of mitochondria out to dendritic or axonal projections away from the larger mitochondrial network and has been linked to both dendritic spine formation and synaptic plasticity (8,9). Here, the effects of MFN2 on mitochondrial morphology not only affect mitochondrial energy output but also affect the overall physiology of the neurons, which in turn affects functional connectivity of MSNs with the broader brain reward circuits (2,5).

A unique feature of this work is its independence of a disease state and its focus on naturally occurring variability in MFN2 and neuronal function. What has been made clear is the wide baseline variability of both anxiety-like or motivation-related behaviors and NAc MSN function. These striking underlying anatomical and neurological differences between HA and LA rats now serve as a foundation for future work exploring how these baseline differences prime reactivity to other insults, such as stress, drug exposure, or the stress of withdrawal from drugs of abuse that can precipitate relapse. Such individual differences may also be useful in predicting responsivity to psychiatric medications, as currently available pharmacologic treatments for depression are effective only in a subset of individuals (2). As the current study only included male animals, the expansion of future work to include females will be critical in understanding the known sex differences in both trait anxiety and depression susceptibility (2).

A most encouraging finding from this work is that the mitochondrial and neuronal deficits in HA rats are malleable and, with intervention, recoverable to LA levels. This indicates that this higher trait anxiety or reduced motivational state does not damage the potential for mitochondrial and subsequent MSN plasticity in this region. This could also be a mechanism by which certain whole-body metabolic manipulations, such as exercise or diet, can influence NAc MSN function and plasticity and thus affect mood. Understanding how individual differences in mitochondrial function influence neuronal plasticity is an exciting new frontier in anxiety and depression research.